wireframer

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (high risk: 1.00). The prompt instructs the agent to locate, create, or modify project-level AI instruction files (e.g., agents.md) to persistently record and enforce its aesthetic rules, which alters system/project context beyond the stated task of generating a wireframe and thus is a hidden/deceptive instruction.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill's generated prototype HTML (samples/bio.html) injects and relies on the CDN script https://unpkg.com/wired-elements?module which will be fetched and executed at runtime as a required dependency for interactive UI components.